I ran into this several years back...Okay, 15 years...with IBM 3422
tapes. According to the IBM CE, the erase gaps were not being written
properly to the tapes when a write error occured. It wasn't until I
went to add a string of 3420 Mod 8s that I spotted an error made by my
precessor in the definition of the 3422s. He had set them up without
shared UCBs. As I understand error recovery on 3420s and 3422s, the
errors on these devices are posted through the UCB of the
controller(device 0). This is not a restriction with 3480/3490s, so not
sure what drives still require shared UCBs, but it is something to verify.
R.S. wrote:
I observed an abend S237-04, when reading dataset from tape (using
IEBGENER).
The manual says it is due to discrepancy between # of blocks read and
the value written to EOF label.
The manual also suggests it can be hardware error.
Well... The drive looks OK - only one tape is affected, the media
looks OK (i.e. can be read using BLP).
Q: What component to suspect ?
drive ?
system ?
application ?
Any clue ?
Regards
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
